Method of fabricating a highly reliable gate oxide

BACKGROUND OF THE INVENTION

The present invention relates generally to semiconductor devices and, in particular, to ultra-thin gate oxide memory devices.

BRIEF SUMMARY OF THE INVENTION

A conventional metal-oxide-semiconductor (MOS) device is illustrated in FIG. <b>1</b>. The device generally includes a gate electrode <b>20</b> which acts as a conductor to which an input signal is typically applied via a gate terminal (not shown). Conventionally doped active areas <b>14</b> and <b>16</b> are formed within the semiconductor substrate <b>10</b> and act as source and drain regions <b>14</b> and <b>16</b>, respectively. A channel region <b>12</b> is formed in the semiconductor substrate <b>10</b> beneath the gate electrode <b>20</b> and separates the source/drain regions <b>14</b>, <b>16</b>. The channel region <b>12</b> is typically doped with a dopant opposite to that of the doped source/drain regions <b>14</b>, <b>16</b>. The gate electrode <b>20</b> is separated from the semiconductor substrate <b>10</b> by an insulating gate oxide layer <b>18</b>, which is typically an oxide of silicon, for example silicon dioxide (SiO₂). The gate oxide layer <b>18</b> prevents current from flowing between the gate electrode <b>20</b> and the semiconductor source region <b>14</b>, the drain region <b>16</b> and/or the channel region <b>12</b>.

When an input voltage is applied to the gate electrode <b>20</b>, a transverse electrical field is set up in the channel region <b>12</b>. By varying the transverse electrical field, the conductance of the channel region <b>12</b> between the source region <b>14</b> and the drain region <b>16</b> is modulated. This way, an electric field controls the current flow through the channel region <b>12</b>. This type of device is commonly known as an MOS field-effect-transistor (MOSFET).

The growth of the gate oxide layer, such as the gate oxide layer <b>18</b> of FIG. 1, is a critical step in manufacturing miniaturized semiconductor devices. Thin gate oxide layers free of defects and of high quality without contamination are essential for proper device operation, especially when current design rules demand gate oxide layers with thicknesses of less than 15 Angstroms, and even less than 10 Angstroms. To obtain high-quality gate oxide layers, the surface of the active area of the device is typically treated with a wet etch to remove any residual oxide. The gate oxide is then grown slowly, typically through dry oxidation in a chlorine ambient. At this point, it is extremely important to carefully control the growth of the gate oxide because the thickness and uniformity of the gate oxide layer can significantly impact the overall operation of the device formed. Because the drain current in a MOS device is inversely proportional to the thickness of the gate oxide, it is desirable to make the gate oxide as thin as possible while taking into account the oxide breakdown and reliability considerations of the process. Furthermore, the use of silicon dioxide for gate oxide layers thinner than 20 Angstroms poses various problems, one of them being the leakage current caused by direct tunneling, which further affects the operation of the device.

High-dielectric constant insulating materials have been proposed as gate oxide layers, but with limited results. FIG. 2 illustrates a high-dielectric constant insulating layer <b>19</b> formed between the gate electrode <b>20</b> and the semiconductor substrate <b>10</b>. Conventional high-dielectric constant insulating materials such as tantalum oxide (Ta₂O₅), titanium oxide (TiO₂) or barium oxide (BaO), for example, are not thermally stable when in direct contact with a silicon substrate. Accordingly, these high-dielectric constant insulating materials require a diffusion barrier layer <b>21</b> (FIG. 2) at the interface with the silicon substrate, the formation of which adds process complexity.

Furthermore, using a diffusion barrier layer defeats the purposes of using a high-dielectric constant insulating material because the gate capacitance is decreased rather than increased. If the gate structure of FIG. 2 is viewed as a series of stacked capacitors <b>25</b> (FIG. <b>3</b>), which has layers of thicknesses comparable to those of the gate structure of FIG. 2, then, a first capacitor C₁(FIG. 3) corresponds to the high-dielectric constant insulating layer <b>19</b> and a second capacitor C₂(FIG. 3) corresponds to the diffusion barrier layer <b>21</b>. The diffusion barrier layer <b>21</b> (FIG. 2) acts as a series capacitor the addition of which decreases the capacitance of the gate electrode <b>20</b>. The capacitance of the first capacitor C₁is larger than the capacitance of the second capacitor C₂and, thus, voltage V₁which occurs across the first capacitor C₁is smaller than voltage V₂which occurs across the second capacitor C₂. As a result, the applied voltage V that occurs across the series capacitors <b>25</b>, that is the sum of V₁and V₂, appears mostly across the diffusion barrier layer <b>21</b> rather than across the high-dielectric constant insulating layer <b>19</b>.

Accordingly, there is a need for an improved memory device which eliminates the problems posed by the use of a conventional high-dielectric constant insulating materials as gate oxide layers. There is also a need for an improved ultra-thin gate oxide layer which is thermally stable when in contact with silicon and which is resistive to impurity diffusion, and a novel method for its fabrication. A memory device with a minimal voltage drop across the gate electrode is also desirable, as well as a method of forming such a memory device.

SUMMARY OF THE INVENTION

The present invention provides an ultra-thin gate oxide layer of hafnium oxide (HfO₂) as a thin medium-dielectric constant gate insulating layer. The ultra-thin gate oxide layer of hafnium oxide (HfO₂) is formed by a two-step process: (1) a thin hafnium (Hf) film is formed by thermal evaporation at a low substrate temperature, after which (2) the thin hafnium film is radically oxidized using a krypton/oxygen (Kr/O₂) high-density plasma to form the ultra-thin gate oxide layer of hafnium oxide (HfO₂). The ultra-thin gate oxide layer of hafnium oxide (HfO₂) formed by the method of the present invention is thermally stable in contact with silicon and is resistive to impurity diffusion at the HfO₂/silicon interface. The formation of the ultra-thin gate oxide layer of hafnium oxide (HfO₂) eliminates the need for a diffusion barrier layer, allows thickness uniformity of the field oxide on the isolation regions and, more importantly, preserves the atomically smooth surface of the silicon substrate.

These and other advantages and features of the invention will be more clearly understood from the following detailed description which is provided in connection with the accompanying drawings.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 illustrates a schematic cross-sectional view of a portion of a conventional MOS device formed in accordance with a method of the prior art.

FIG. 2 illustrates a schematic cross-sectional view of a portion of a conventional MOS device formed in accordance with another method of the prior art.

FIG. 3 is a schematic view of a series of capacitors corresponding to stacked layers of the structure of FIG. <b>2</b>.

FIG. 4 illustrates a schematic cross-sectional view of a portion of a memory device formed according to a method of the present invention.

FIG. 5 illustrates a schematic cross-sectional view of the memory device of FIG. 4 at a stage of processing subsequent to that shown in FIG. <b>4</b>.

FIG. 6 illustrates a schematic cross-sectional view of the memory device of FIG. 4 at a stage of processing subsequent to that shown in FIG. <b>5</b>.

FIG. 7 illustrates a schematic cross-sectional view of the memory device of FIG. 4 at a stage of processing subsequent to that shown in FIG. <b>6</b>.

FIG. 8 illustrates a schematic cross-sectional view of the memory device of FIG. 4 at a stage of processing subsequent to that shown in FIG. <b>7</b>.

FIG. 9 illustrates a schematic cross-sectional view of the memory device of FIG. 4 at a stage of processing subsequent to that shown in FIG. <b>8</b>.

FIG. 10 illustrates a schematic cross-sectional view of the memory device of FIG. 4 at a stage of processing subsequent to that shown in FIG. <b>9</b>.

FIG. 11 illustrates a schematic cross-sectional view of the memory device of FIG. 4 at a stage of processing subsequent to that shown in FIG. <b>10</b>.

FIG. 12 illustrates a schematic cross-sectional view of the memory device of FIG. 4 at a stage of processing subsequent to that shown in FIG. <b>11</b>.

FIG. 13 illustrates a schematic cross-sectional view of the memory device of FIG. 4 at a stage of processing subsequent to that shown in FIG. <b>12</b>.

FIG. 14 illustrates a computer system having a memory cell with a gate stack including the ultra-thin gate oxide layer of hafnium oxide (HfO₂) formed according to the present invention.

DETAILED DESCRIPTION OF THE INVENTION

In the following detailed description, reference is made to various specific embodiments in which the invention may be practiced. These embodiments are described with sufficient detail to enable those skilled in the art to practice the invention, and it is to be understood that other embodiments may be employed, and that various structural, logical, and electrical changes may be made without departing from the spirit or scope of the invention.

The term “substrate” used in the following description may include any semiconductor-based structure that has an exposed substrate surface. Structure must be understood to include silicon-on insulator (SOI), silicon-on sapphire (SOS), doped and undoped semiconductors, epitaxial layers of silicon supported by a base semiconductor foundation, and other semiconductor structures. When reference is made to a substrate or wafer in the following description, previous process steps may have been utilized to form regions or junctions in or over the base semiconductor or foundation.

The term “hafnium” is intended to include not only elemental hafnium, but hafnium with other trace metals or in various alloyed combinations with other metals as known in the semiconductor industry, as long as such hafnium alloy is conductive, and as long as the physical and electrical properties of the hafnium remain unchanged. Similarly, the term “hafnium oxide” is intended to include not only elemental hafnium oxide, but hafnium oxide with other trace metals or in various alloyed combinations with other metals as known in the semiconductor industry, as long as the physical and electrical properties of the hafnium oxide remain unaltered.

The present invention provides an improved memory device comprising gate structures including an ultra-thin medium-dielectric constant gate insulator formed of hafnium oxide. The ultra-thin medium-dielectric constant gate insulator of the invention reduces the voltage drop across the gate electrode, preserves the smoothness of the substrate surface and permits effective transistor operations without the need of a diffusion barrier layer.

Referring now to the drawings, where like elements are designated by like reference numerals, FIGS. 4-13 illustrate the formation of one embodiment of an improved memory device <b>100</b> (FIG. 13) of the present invention. FIG. 4 illustrates a semiconductor substrate <b>50</b> having a well <b>52</b>, which is typically doped to a predetermined conductivity, for example P-type or N-type, depending on whether NMOS or PMOS transistors will be formed therein. For the purposes of the present invention, the well <b>52</b> is considered to be a P-well formed by implanting P-type impurity atoms, such as boron, for example, by known methods of the art. In this application, the semiconductor substrate <b>50</b> will be referred to as a P-type silicon substrate <b>50</b>, but it must be understood that the substrate need not be silicon-based. Thus, the present invention has equal applicability to other semiconductor substrates, such as, for example, silicon-germanium, germanium, silicon-on-saphire, or gallium-arsenide substrates, among others.

The structure of FIG. 4 further includes isolation regions <b>53</b>, which, as known in the art, may be field oxide (FOX) regions formed by a shallow trenches for isolation (STI) process, for example. In an exemplary embodiment, the isolation regions <b>53</b> are shallow trenches for isolation filled with high density plasma (HDP) oxide, a material which has a high ability to effectively fill narrow trenches. Alternatively, an insulating layer (not shown) formed of an oxide or of silicon nitride, for example, may be formed on the trench sidewalls, prior to filling the trenches with the isolation dielectric, to aid in smoothing out the corners in the bottom of the trenches and to reduce the amount of stress in the dielectric used to later fill in the trenches.

Subsequent to the formation of the isolation regions <b>53</b> (FIG. <b>4</b>), an ultra-thin layer of insulating material <b>60</b> of hafnium (Hf) is formed over the silicon substrate <b>50</b> including the isolation regions <b>53</b>, as shown in FIG. <b>5</b>. Hafnium is the preferred material for the insulating material <b>60</b> because, as it will be explained in more detail below, hafnium forms the most stable oxide with the highest heat of formation (ΔHf=271 Kcal/mol) among the elements of group IVA of the periodic table (Ti, Zr, Hf). Hafnium can also reduce the native silicon dioxide (SiO₂) layer to form hafnium oxide (HfO₂). Hafnium is further preferred because hafnium oxide (HfO₂) has a medium-dielectric constant (k) of about 30 with a bandgap of 5.68 eV. Hafnium oxide (HfO₂) is very resistive to impurity diffusion and intermixing at the interface HfO₂/silicon because of its high density, which is of about 9.69 g/cm³.

In a preferred embodiment of the invention, the ultra-thin hafnium layer <b>60</b> (FIG. 5) is formed by thermal evaporation, for example by electron-beam evaporation, using an ultra-high purity Hf metal slug, of about 99,9999% purity, and at a low temperature substrate of about 150° C. to about 200° C. The ultra-thin hafnium layer <b>60</b> is formed over the whole surface of the silicon substrate <b>50</b>, including the isolation regions <b>53</b>, to a thickness of about 10 Angstroms to about 100 Angstroms, more preferably of about 10 Angstroms to about 50 Angstroms.

The formation of the ultra-thin hafnium layer <b>60</b> (FIG. 5) by a method of the present invention has advantages over conventional deposition techniques. Conventional hafnium metal deposition processes of the prior art include reactive sputtering, such as magnetron sputtering in an argon/oxygen ambient or in an argon ambient, during which ion bombardment damages the surface of the silicon substrate. In contrast, the formation of the ultra-thin hafnium layer <b>60</b> by thermal evaporation preserves the original atomically smooth surface of the silicon substrate <b>50</b>. Preservation of the smoothness of a silicon surface is an important property in the semiconductor industry, as the roughness of the substrate surface has been recently related to the leakage current. In a recent article, <i>The electronic structure of the atomic scale of ultrathin gate oxides</i>, Nature, Vol. 399, 1999, pp. 753-761, Muller et al. have emphasized that a fundamental characteristic of an ultra-thin gate oxide layer is the ability of the oxide to control the roughness of the oxide/silicon interface at an atomic scale. Muller et al. have demonstrated that the leakage current through a 10 Angstroms oxide increases by about a factor of 10 for every 0.1 Angstroms increase in the root-mean-square (rms) roughness. The leakage current in conjunction with the sub-threshold leakage are the most important characteristics of MOSFET devices. Thus, the method of the present invention for forming the ultra-thin hafnium layer <b>60</b> of FIG. 5 avoids damage of the silicon substrate <b>50</b>, preserves the originally atomically smooth surface of the silicon substrate <b>50</b> and, therefore, reduces the leakage current.

Subsequent to the formation of the ultra-thin hafnium layer <b>60</b> (FIG. <b>5</b>), an ultra-thin hafnium gate oxide (HfO₂) layer <b>62</b> (FIG. 6) is formed by oxidation of the ultra-thin hafnium layer <b>60</b>. In a preferred embodiment of the invention, the ultra-thin hafnium layer <b>60</b> is oxidized at low temperatures by atomic oxygen generated in a high-density krypton (Kr) plasma. Precise details of a silicon oxidation technique, which can be employed for the hafnium oxidation of the present invention, were given recently by Saito et al., in <i>High</i>-<i>Intergity Silcion Oxide Grown at Low</i>-<i>Temperature by Atomic Oxygen Generated in High</i>-<i>Density Krypton Plasma</i>, Extended Abstracts of the 1999 Int'l Conference on Solid State Devices and Materials, (Tokyo, 1999), pp. 152-53, the disclosure of which is incorporated by reference herein. However, a brief summary of how an ultra-thin gate oxide layer is formed by atomic oxygen generated in a high-density krypton (Kr) plasma is believed to be helpful to the understanding of the present invention.

As Saito et al. have reported, silicon dioxide films were grown by direct oxidation of a silicon surface at about 400° C. and at low electron temperature (below 1.3 eV), low ion bombardment energy (less than 7 eV) and high plasma density (above 10¹²/cm³). A mixed gas of about 3% oxygen in krypton (O₂/Kr) was used at a pressure of 1 Torr and a microwave density of about 5 W/cm². The typical growth rate of the silicon oxide was 14 Angstroms for about 10 minutes. In <i>Low</i>-<i>Temperature Growth of High</i>-<i>Intergity Silicon Oxide Films by Oxygen Radial Generated in High</i>-<i>Density Krypton Plasma</i>, IEEE Tech. Dig., 1999, pp. 249-252, Hirayama et al. further analyzed film growth in O₂/Kr plasma and demonstrated that the oxidation rate of O₂/Kr plasma is faster than the oxidation rate of an O₂/He plasma. Hirayama et al. also showed that the growth rate of O₂/Kr mixed plasma is higher than the growth rate of thermally grown silicon oxide films at 1000° C.

In accordance with the findings of Saito et al. and Hirayama et al., the ultra-thin hafnium layer <b>60</b> (FIG. 5) undergoes oxidation in a high-density microwave 3% oxygen/krypton (O₂/Kr) plasma at a temperature of about 400° C. to form the ultra-thin hafnium gate oxide (HfO₂) layer <b>62</b> of FIG. <b>6</b>. The ultra-thin hafnium gate oxide (HfO₂) formed by the method of the present invention has a high growth rate, high dielectric strength, and low interface trap and bulk charge. Furthermore, according to the findings of Saito et al., the ultra-thin hafnium gate oxide (HfO₂) can displace native thermally grown silicon dioxide.

Another advantage of using the oxygen/krypton (O₂/Kr) plasma for the oxidation of the ultra-thin hafnium layer <b>60</b> of FIG. 5 is that the thickness variation of the ultra-thin hafnium gate oxide (HfO₂) layer <b>62</b> (FIG. 6) at the edges of the isolation regions <b>53</b> (FIG. 6) is dramatically improved compared with the thickness variation of conventional thermal oxidation processes. This is because the oxide growth in an oxygen/krypton (O₂/Kr) plasma does not depend on surface orientation and oxygen radicals can penetrate any place, even very complicated structures, due to their very small atomic radius. These findings have been reported by Saito et al. in <i>Advantage of Radical Oxidation for Improving Reliability of Ultra</i>-<i>Thin Gate Oxide</i>, VLSI Tech. Dig., (2000 Symposium on VLSI Technology Digest of Technical Papers), 2000, pp. 176-77, the disclosure of which is incorporated by reference herein. Accordingly, the radical oxidation of the ultra-thin hafnium layer <b>60</b> of FIG. 5 may be conducted on a silicon surface of any orientation, for example on a silicon surface of <<b>111</b>> orientation, and the invention is not limited to a silicon surface of <<b>100</b>> orientation which is characteristic to the thermal oxidation processes of the prior art.

After the formation of the ultra-thin hafnium gate oxide (HfO₂) layer <b>62</b> (FIG. <b>6</b>), a conductive layer <b>64</b> is formed over the ultra-thin hafnium gate oxide (HfO₂) layer <b>62</b>, as shown in FIG. <b>7</b>. The conductive layer <b>64</b> may be formed, for example, of doped polysilicon, metals, metal silicides, conductive metal oxides, or combinations of conductive materials, for example, a refractory metal silicide layer overlying a doped polysilicon layer. In any event, the ultra-thin hafnium gate oxide (HfO₂) layer <b>62</b> acts as an impurity diffusion barrier layer which allows the fabrication of the gate structures <b>70</b> (FIG. 11) without the depletion effect.

In an exemplary embodiment of the invention, the conductive layer <b>64</b> is formed of polysilicon, which may be deposited over the ultra-thin hafnium gate oxide (HfO₂) layer <b>62</b> via LPCVD procedures, at a temperature of about 300° C. to about 700° C., and to a thickness of about 100 Angstroms to about 2,000 Angstroms. The polysilicon layer <b>64</b> may be doped in situ during deposition with arsine or phosphine in a silane ambient, or the polysilicon layer <b>64</b> may be deposited intrinsically and then doped by ion implantation of arsenic or phosphorous ions. After its deposition, the polysilicon layer <b>64</b> may be planarized by chemical mechanical polishing (CMP), for example, but other suitable methods could be used also, as desired. In any event, the formation of the ultra-thin hafnium gate oxide (HfO₂) layer <b>62</b> by the method of the present invention eliminates the need for an additional barrier layer between the ultra-thin hafnium gate oxide (HfO₂) layer <b>62</b> and the polysilicon layer <b>64</b> to prevent reaction between the polysilicon and the hafnium oxide.

In an exemplary embodiment of the invention, silicide regions <b>65</b> may be formed over the polysilicon layer <b>64</b>, as illustrated in FIG. <b>8</b>. If the silicide regions are desired, a layer of metal capable of forming a silicide (not shown) may be deposited over the polysilicon layer <b>64</b> (FIG. 7) by sputtering by RF or DC or by other similar methods such as CVD, to a thickness of about 200 Angstroms to about 500 Angstroms. Subsequent to the deposition of the metal capable of forming a silicide, the silicon substrate <b>50</b> undergoes a rapid thermal anneal (RTA), typically for about 10 to 60 seconds, using a nitrogen ambient, at about 600° C. to about 850° C. so that the metal in direct contact with the polysilicon layer <b>64</b> is converted to its silicide. As shown in FIG. 8, silicide layer <b>65</b> forms a conductive region on top of the polysilicon layer <b>64</b>. Preferably, the refractory metal has low resistance and low resistivity as a silicide. However, the refractory metal silicide may comprise any refractory metal, including but not limiting to titanium, cobalt, tungsten, tantalum, molybdenum, and platinum.

Although the following processing steps for the completion of the gate stacks <b>70</b> (FIG. 11) will refer to and illustrate the silicide layer <b>65</b> formed over the polysilicon layer <b>64</b>, it must be understood that the present invention is not limited to this embodiment, and other embodiments such as the formation of gate stacks without a silicide layer on the polysilicon gate, are also contemplated. A protective cap material <b>66</b> is formed over the silicide layer <b>65</b> as shown in FIG. <b>8</b>. The cap material may be formed of silicon dielectrics such as silicon nitride or silicon oxide, but TEOS or carbides may be used also. The cap material <b>66</b> may be formed via PECVD deposition procedures, for example, at a temperature between about 300° C. to about 600° C., to a thickness of about 500 Angstroms to about 2,000 Angstroms.

Next, the structure of FIG. 8 is patterned using a photoresist layer <b>67</b> (FIG. 9) formed over the cap material <b>66</b> to a thickness of about 1,000 Angstroms to about 10,000 Angstroms. The photoresist layer <b>67</b> is patterned with a mask (not shown) and the ultra-thin hafnium gate oxide (HfO₂) layer <b>62</b>, the polysilicon layer <b>64</b>, the silicide layer <b>65</b> and the cap material <b>66</b> are anisotropically etched through the patterned photoresist to obtain a plurality of polysilicon gates <b>70</b><i>a </i>having on top portions <b>67</b><i>a </i>of the photoresist layer, as shown in FIG. <b>10</b>. Subsequent to the formation of the polysilicon gates <b>70</b><i>a</i>, the top portions <b>67</b><i>a </i>of the photoresist layer are removed by conventional techniques, such as oxygen plasma, for example, or by flooding the silicon substrate <b>50</b> with UV irradiation to degrade portions <b>67</b><i>a </i>of the photoresist layer and obtain gate stacks <b>70</b> of FIG. <b>11</b>.

As illustrated in FIG. 11, each of the gate stacks <b>70</b> comprises the ultra-thin hafnium gate oxide (HfO₂) layer <b>62</b>, the polysilicon layer <b>64</b>, the silicide layer <b>65</b> and the protective cap material <b>66</b>. The gate stacks <b>70</b> may now be used in a conventional implant process where the gate stacks are needed to mask the dopant implantation of source/ drain regions <b>82</b> (FIG. 12) of the adjacent transistors defined by the gate stacks. As such, FIG. 11 illustrates the formation of lightly doped drain (LDD) regions <b>72</b> in the silicon substrate <b>50</b>. As known in the art, the LDD regions <b>72</b> may be formed by implanting low dosages of conductivity-altering dopants with an LDD mask (not shown) or using the gate stacks <b>70</b> as an implantation mask. Thus, in the P-well <b>52</b> of the silicon substrate <b>50</b>, N-type dopants such as arsenic or phosphorous may be implanted at a low energy dose, for example of about 1×10¹⁵atoms/cm², using the gate stacks <b>70</b> as a mask.

The next step in the process flow is the formation of spacers <b>68</b> illustrated in FIG. <b>12</b>. Spacers <b>68</b> may be formed, for example, by depositing a silicon nitride film or a silicon oxide material over the structure of FIG. <b>11</b> and then anisotropically etching with an RIE plasma to form the spacers <b>68</b> on each of the sidewalls of the gate stacks <b>70</b>.

The gate stacks <b>70</b> protected by spacers <b>68</b> can now undergo conventional processing steps for the formation of source/drain regions in the silicon substrate <b>50</b>. As such, using the gate stacks <b>70</b> as an implantation mask, heavily doped source/drain regions <b>82</b> are formed in the uncovered portions of the silicon substrate <b>50</b> via an ion implantation procedure performed at an energy of about 50 KeV to about 70 KeV, and at a dose of about 2×10¹⁵to about 5×10¹⁵atoms/cm², using arsenic or phosphorous, for example. Thermal annealing for activation may be carried out in a nitrogen atmosphere at about 1000° C. for about 10 minutes.

Conventional processing steps can now be carried out to form conductors <b>89</b> and/or capacitors <b>99</b> with associated conductive plugs <b>99</b><i>a</i>, all illustrated in FIG. 13, and to complete the formation of a DRAM memory device <b>100</b>. For this, an insulating layer <b>97</b> is formed over the structure of FIG. <b>12</b> and contact openings (not shown) are created through the insulating layer <b>97</b>. The insulating layer <b>97</b>, may include borophosphosilicate glass (BPSG) or phosphosilicate glass (PSG), for example, and may be formed by normal CVD or LPCVD processes to a thickness of about 1,000 Angstroms to about 5,000 Angstroms. To create contact openings through the insulating layer <b>97</b>, a photoresist material (not shown) is deposited and patterned using conventional photolithography steps and the insulating layer <b>97</b> is then etched, by a wet etch for example, so that the contact openings extend to source/drain regions <b>82</b>. Conductors <b>89</b> and/or associated conductive plugs <b>99</b><i>a </i>of capacitors <b>99</b> are then formed by known metallization processes.

The gate stacks <b>70</b> including the ultra-thin hafnium gate oxide (HfO₂) layer <b>62</b> formed in accordance with embodiments of the present invention could be used in any integrated circuit structure such as in a processor-based system <b>400</b> (FIG. 14) which includes a memory circuit <b>448</b>, for example a memory module containing a plurality of DRAM memory devices <b>100</b> having gate stacks comprising the ultra-thin hafnium gate oxide (HfO₂) layer <b>62</b> formed according to the present invention. A processor system, such as a computer system, generally comprises a central processing unit (CPU) <b>444</b>, such as a microprocessor, a digital signal processor, or other programmable digital logic devices, which communicates with an input/output (I/O) device <b>446</b> over a bus <b>452</b>. The memory <b>448</b> communicates with the system over bus <b>452</b>. The processing unit <b>444</b> and other devices illustrated in FIG. 14 may also contain circuits having transistors with gate stacks comprising the ultra-thin hafnium gate oxide (HfO₂) layer <b>62</b> formed according to the present invention.

Although the invention has been illustrated for a DRAM memory device, such as the DRAM memory device <b>100</b> (FIG. <b>13</b>), fabricated on a p-type substrate, the invention could also be fabricated on an n-type substrate, as well-known in the art. This, of course, will change the doping and conductivity of the operative layers in the fabricated device.

The above description and drawings are only to be considered illustrative of exemplary embodiments which achieve the features and advantages of the present invention. Modification and substitutions to specific process conditions and structures can be made without departing from the spirit and scope of the present invention. Accordingly, the invention is not to be considered as being limited by the foregoing description and drawings, but is only limited by the scope of the appended claims.